About half of exposures have no symptoms at all, about half of the remaining ones have mild cold symptoms (fever, sore throat, dry cough), and then a further 30% of the remainder might have nasty symptoms (deep cough, walking pneumonia) that don't require hospitalization. The remaining 20% of symptomatic cases (10% of total) get pneumonia that's bad enough to require hospitalization, and half of those require ventilators and ICU beds. With good medical care and no pre-existing conditions about 90% of those patients recover, without medical care virtually all of them die.

Untreated, it's basically like playing Russian Roulette. 5 out of 6 times, nothing bad happens. 1 out of 6, you die. The disparity in outcomes is probably what makes this so insidious: a lot of folks look at the numeric outcomes for healthy people in the presence of ventilators and ICUs and reasonably think "Oh, I have a 1 in 500 chance of dying, it's no big deal". Medical professionals look at the exponential infection curve, the capacity to overwhelm health care systems, and the number of patients requiring hospitalization and think "this is the scariest disease I've ever seen".

One thing missing from your analysis is that the group you end up in is greatly impacted by age and current health.

Right, although the news out of Italy is that young healthy adults are ending up in the "bad pneumonia, need hospitalization" bucket nearly (but not quite) as much as older patients. They just have a much better prognosis with hospital interventions, because once they're ventilated or on oxygen for a while their immune system can reliably clear the virus, while with 80+ year olds or folks with pre-existing conditions it starts running amuck and causes ARDS or multiple organ failure. Overload the hospital system and the death rate jumps for healthy young adults as well - this happened in Wuhan and is now happening in Italy and Iran, and is why you see videos on the Internet of middle-aged adults dead in their bed from coronavirus.

Your analysis is suffering from survivorship bias -- Italy still has no idea how many infections they actually have.

South Korea is a better dataset because they have tested so many people.

Could you share the source for this? It's very useful for understanding the situation.

Some reports are showing that a little under half of those infected get no symptoms at all. Otherwise it can be anything from mild cold symptoms all the way to deadly pneumonia. That's from what I've been reading in the UK anyway.

Can someone be tested after recovery to see if they had it? Would that person be a possible carrier?